Results tables display the gene-wise counts from each replicate in a given contrast, followed by the fold change value of each gene and the associated p-value (adjusted) of the negative binomial hypothesis test conducted by DESeq2. Lower p-values indicate a lower probability of the null hypothesis that counts between the two conditions are derived from the same distributional parameters. Furthermore, common gene names from any uploaded gene names table will appear in a column alongside standard gene IDs for the genes that aren’t associated with a common name. In addition to pdf outputs, these tables can be saved as html widgets with sorting, searching, and page size customization features.
The PCA plot displays loadings of the first two principal components for each sample/biological replicate in the experimental design. Colored by experimental condition, the points of the PCA plot provide a visualization of clustering amongst the samples, both within conditions and across conditions.
The Intra-Condition Scatter Plots display log2 counts between pairs of biological replicates within each condition of the experimental design. Any counts in the data below 1 are replaced with a value of 1 to simplify the log2 transformation. All replicate pairs in each condition are displayed.
The Mean Reads Scatter Plots display average log2 counts across biological replicates of experimental condition pairs for a provided contrast. Average counts with a value of 0 are assigned the value -4 following the log2 transformation. Guidelines are added to assist with visualization. Statistically significant genes are less transparent than insignificant genes, and they are colored blue in default plots. Axes are scaled to more closely resemble the log2 transformation.Unlabeled tick marks, therefore, do not always represent whole number intervals between labeled tick marks. In addition to pdf outputs, Mean Reads Scatter Plots can be saved as html widgets with draggable zooming, panning, and hover text features. To reset the scatter plot axes within the widget, the home button in the top right corner of the widget can be pressed.
The mean reads scatter plots can be customized by changing the p-value and fold change thresholds for distinguishing statistically significant genes. Furthermore, upper and lower transparency thresholds can be set for distinguishing statistical significance. The lower threshold corresponds to insignificant genes. The points of the plot can also be colored and sized according to their gene classes as specified by the gene table. If no gene table has been selected/uploaded, only the p-value, fold change, and transparency thresholds will be customizable. If a gene table has been selected/uploaded, a class parameters csv can be selected from the working directory. If used, only points/genes corresponding to classes in the table will be plotted. If no class parameters csv is selected, every class will be automatically colored from a list of 15 colors, and every point will be sized with a cex value of 0.3. Insignificant genes will be colored grey if the customize by significance feature is set to TRUE.
The Standard MA Plots are built from the DESeq2 package and display the fold change of a gene over its mean counts value (normalized) for a provided contrast between experimental conditions. Genes with a statistically significant p-value (p < 0.05) are colored in blue.
The Complete heatmap shows log2 counts across all samples for any genes above a certain mean count threshold of three (meaning an average of eight counts across all samples). Using the package heatmaply, an interactive html widget is rendered with draggable zooming, panning, and hover text features. To reset the heatmap axes, the home button in the top right corner of the widget can be pressed. Darker blue cells indicate lower log2 counts values, while darker red cells indicate higher log2 counts values. Furthermore, rows (genes) are clustered using the complete hierarchical clustering method in R (hclust) with eudclidean distances.
The Class-Separated: All Classes heatmap relies on the gene table to map all genes with an associated class. Such genes are grouped alphabetically by class and clustered hierarchically within their respective group. Hover text indicates the class assigned to each gene.
The Class-Separated: Selected Classes heatmap produces subplots for each class listed in the associated yaml parameter entry. Heatmaps for each listed class are scaled according to the limits of the complete heatmap, although cell sizes are scaled according to the number of genes associated with each class.